Variation in Fertilization Rate in the Tropical Reef Fish, Halichoeres bivattatus: Correlates and Implications

6 Pages·1991·2.5 MB·English

by  C W Petersen

Reference: Biol Bull 181: 232-237. (October. 1991)
Variation in Fertilization Rate in the Tropical
Reef Fish, Halichoeres bivattatus:
Correlates and Implications
CHRISTOPHER W. PETERSEN
Smithsonian TropicalResearch Institute. Box 2072, Balboa, Panama
Abstract. Fertilization rates were estimated for natural the adaptive significance ofmorphological and behavioral
spawns in the tropical wrasse, Halichoeres bivattatus. differences between and within the sexes.
Fertilization rate averaged 88%, but varied with both sea Thewrasses have reproductivebehaviorstypical offree-
conditions and with the addition of males (streakers) to spawning tropical reeffishes, with spawning occurring at
pairspawns. Asseaconditionsbecame rougher, the mean downcurrent sites of reefs. Spawning consists of one to
fertilization rate for a day decreased. This effect was due overten malesandafemalerushingupward0.2 toseveral
to the two days with the roughest sea conditions and the meters toward the surface and releasing gametes into the
lowest mean fertilization rates; there was noobvioustrend watercolumn attheapex ofthe spawningrush. Although
when thesetwodayswereexcluded from theanalysis. On pastworkershaveproposed manyalternative mechanisms
a given day, pair spawns, with a single male and female, for the evolution of reproductive behaviors in tropical
had fertilization rates approximately 7% lower than reeffishes, one fundamental factorhas been almost com-
spawnswhere 1-2 streakersjoined the pairspawn. These pletely ignored: theeffect offertilization rateon individual
results suggest that variation in fertilization rate must be reproductive success and spawning behavior (for excep-
considered asa potential selective force in shaping repro- tions see Shapiro, 1989; Petersen, 1991). In this paper, I
ductive behavior in fishes with external fertilization of provide measurementsforfertilization rateinthetropical
pelagic eggs. wrasse, Halichoeres bivattatus, examine patterns and the
potential causes ofvariation in fertilization rate, and dis-
Introduction cuss the role that fertilization rate may play in the evo-
lution ofreproductive behavior in this species.
The social and mating systems of tropical reef fishes, Selection on males has been generally thought to act
especiallythemorecolorfulandabundantspecieslikethe such that individuals produce enough sperm to fertilize
wrasses (Pisces: Labridae), have been the focus ofa large all ofthe eggsofa female, implying that investigations of
body ofresearch (Robertson, 1972, 1981; Robertson and variation in fertilization rate would be uninteresting.
Choat, 1974; Robertsonand Hoffman, 1977; Warnerand However, as the amount of sperm released in a spawn
Robertson, 1978; Thresher, 1979; Warnerand Hoffman, increases, the number of additional fertilizations for a
1980a, b; Moyer and Yogo, 1982; Tribble, 1982; Kuwa- given numberofsperm released should decrease, reducing
mura, 1984; Warner, 1984a, b, 1987, 1988; Nemtzov, the benefit for producing more sperm (Petersen, 1991).
1985; Hoffman, 1985; Ross, 1986; Lejeune, 1987; Victor, There is no reason toexpect that the benefitsto males for
1987; Baird, 1988). Much ofthis work has involved doc- producing additional sperm will outweigh the costs of
umenting spawning behavior, quantifying the reproduc- sperm production untilalleggsarefertilized. Underthese
tive success ofvarious mating tactics, and speculating on circumstances, theselected level ofsperm production may
not be enough to fertilize all available eggs, and the fer-
tilization rate should be less than 100% (Petersen, 1991).
RPreecseeinvtedadd1rAespsn:lC1o9l9l1e;geacocfepttheedAt2l4anMtiacy, 1190951.Eden St., Bar Harbor, The extent to which fertilization rate is important in
Maine04609-1198. discussions ofmating-system evolution depends on how
232
FERTILIZATION RATE IN HALICHOERES 233
much variance exists in fertilization rate and how much streaking pair spawns and, on one occasion, approxi-
ofthat variance ispredictable. Given that fertilization rate mately 10 IP malestook part in agroupspawn. Both pair
may not always be 100%, it is still not clear whether fer- and groupspawningappearto becommon in thisspecies
tilization rateplaysasignificant role in determining male (Warnerand Robertson, 1978). Alldata reported hereare
and female mating strategies in organisms that have ex- from pair spawns with or without streakers.
ternal fertilization ofpelagic eggs.
In many tropical reef fishes, there is more than one Materials and Methods
male mating behavior within a population (e.g.. Warner
et a/.. 1975; Robertson and Warner, 1978; Warner and Data were collected during February 1990 along the
Robertson, 1978; Thresher, 1984; Petersen, 1990). northeast coast of St. Croix, Virgin Islands, and during
Spawningsitesmaybeoccupiedbyeitheralargeterritorial March 1990 in the San Bias Islands of Panama. At St.
male who mates singly with the female (pair spawning) Croix, collections ofeggswere made in shallow water(1-
or by a group ofsmaller males who mate together with 4 m depth) while snorkeling behind the barrier reefap-
the female (group spawning). In addition, small non-ter- proximately 1 km west of Tague Bay. In the San Bias,
ritorial males are known to rush in andjoin pair spawns collections were made while snorkeling at 1-3 m depth
atthe moment ofgamete release, a matingbehaviorcalled at Smithsoniantupo reef adjacent to the Smithsonian
streaking (Warner el al.. 1975; Thresher, 1984). In many Tropical Research Institute San Bias Field Station.
species, there may be several active spawning locations To collect eggs from natural spawnings, observers po-
m
within the range of an individual female, providing fe- sitioned themselves within 2-4 ofa spawningsite near
maleswithachoiceofconditions forspawning. Theextent thebeginningofthedailyspawningperiod. When aspawn
that maleschoose orcompete forspawninglocationsand was observed, the gamete cloud was marked by a small
females choose either spawning locations or the identity amount offluorescein dye released by the observer near
of the male-role spawners based on fertilization rate is the spawn. After waiting for at least 30 s, the observer
unknown. swept the area around and including the expanded fluo-
rescein cloud for 30 s with a 6-inch brine-shrimp net
(maximum mesh size approximately 100 X 300 nm}.
Reproductive Biology ofHalichoeres bivattatus
Collections ofeggs from another wrasse (Thalassoma bi-
The slippery dick, Halichoeres bivattatus, is the most fasciatum. Petersen et al.. 1992) indicated that harsher-
common Caribbean wrasse found over shallow reefs and textured nylon nets, such as plankton netting, severly re-
seagrass communities. Smaller males and females have duced the fertilization rates while the brine-shrimp net
an initial-phase(IP)drabcoloration; largeterminal-phase had no effect. At the end ofthesweep, the netwasdrawn
(TP) individuals are more brightly colored and are in- away from the fluorescein cloud, the contents ofthe net
variably male (Warner and Robertson, 1978). TP males were transferred to a small plastic bag, and the bag was
may be the result ofsex change, or merely a coloration sealed. For each spawn, the location of the spawn was
change from IP males (Warner and Robertson, 1978). recorded, and the number of streakers that joined the
Sex ratio is only slightly biased towards females (1:1.2), spawn (0-2) was noted.
and IP males outnumber TP males approximately 2.3:1 The next morningthecontentsofeach bagwere filtered
(Warner and Robertson. 1978). through 100 nm nylon mesh to collect the eggs, which
Spawning occurs daily in the mid-afternoon over the were then examined under a dissecting microscope and
lee edges ofreefs and over turtle grass (Warner and Rob- scored as fertilized or unfertilized. By counting eggs 16-
ertson, 1978). In pair spawning, gravid females enter the 20 h post-spawning, fertilized, developing eggs could be
territories ofTP males, are approached by the TP male, unambiguously distinguished from undeveloped eggs.
and after one to several rushes by the male toward the Developing eggs contain nearly fully developed larvae,
female both rise quickly 0.3 to over 1 m in the water because hatching iscompleted within 24 h. Undeveloped
column in a fast spawning rush. Gametes are released at eggswere scored as unfertilized. Estimatesoffertilization
the apex ofthis spawning rush, afterwhich both fish dart rate forspawns relied on collections with at least 20 eggs.
back to their normal positions closer to the bottom. The To test whether the 30-s delay before collection was
eggs are fertilized externally as they begin to float away adequatetoallow fertilization tooccur, onthreedayscol-
from the location ofthe spawn. The TP male continues lections were alternated between 30-sdelaysand 60-sde-
tocourtadditional females, and mayspawn over20times lays. A significant increase in fertilization rate in the 60-
in a day (pers. obs.). Streaking occurs when an IP male s sample would suggest that the 30-s samples underesti-
joinsthespawningpairattheapex oftheirspawningrush mate fertilization rate. In addition, collectionsweremade
and presumably releases sperm. Only TP males were ob- at spawninglocationsduringthespawningperiodbut not
servedpairspawningin thisstudy; IPmaleswereobserved immediately after a spawn to verify that eggs collected
234 C. W. PETERSEN
50 i plying that eggs in collections do come largely from the
spawn that was observed immediately prior to sampling.
C 40- There was no evidence that collecting eggs after 30 s
03 detrimentally affected the estimate of fertilization rate.
Q. Eggs collected from spawns after 60 s had virtually iden-
(/D 30-
t=ical fertilPiz=ation rates tospawnscollected after 30 s(F, ,5
0.001, 0.98).
0) 20- Spawning date had a significant effect on fertilization
.0
E rate. Over the 18 dates with at least two collections with
10- 20eggs, fertilization rate varied significantly amongdates
(ANOVA, FI784 = 8.3!,/>< 0.001).
\/A Oneofthecausesofvariation in fertilization rateamong
days appeared to be sea conditions. There wasa negative
20 40 60 80 correlation between thequalitative score ofseacondition
Fertilization rate and the mean fertilization rate for a day, with rougher
witFhig>u2r0e e1.ggs.FrTehqeuelonwceyrdhisoturnibdutfioornthoeff1e0rt%ilciaztaetgioonrireasteissgfiovreanl,lsthpeawlansst d/Va.yasiiehda]vi<ng0.l0o5,wern f=ert1i8lizdaatyiso)n.rTatheiss(aFinga.ly2s,isi\u=sed-0.o4n4l,y
category included twospawnswith 100% fertilizationofcollectedeggs. tshpeawsnpsawwnisthwoiutth ssttrreeaakkienrgs;(rasn=id-e0n.t4i8ca,l/t3[ri.etan,dicd]ex<is0t.e0d5,fonr
= 16). Much ofthe variation in fertilization rates among
after spawns were not in the water column from spawns dates was due to the two dates that had the roughest sea
elsewhere. conditionsandalsohadthelowest mean fertilization rates.
On each date, a qualitative score ofsea conditions was Excluding these two days, there was no significant effect
madeduringthe spawningperiod. Thisscorevaried from of sea conditions on fertilization rate (both types of
>
1 for the calmest conditions to 5 for the roughest condi- spawns, P|Mj,i,d] 0.25).
tions experienced duringthe study. These datawere used Spawns with streakers had significantly higher fertil-
to test fordecreased fertilization rate on days with higher ization rates than pairspawns without streakerscollected
water velocities and increased water mixing. Based on on the same day. To test for changes in fertilization rate
theoretical studies of water turbulence and fertilization with the presence ofstreakers, date ofcollection and the
rate, increasedwatermixingwaspredictedtoreducesperm presence ofstreakers were examined simultaneously in a
ANOVA
concentrations and reduce fertilization rates (Denny and two-way for the subset ofdays with collections
Shibata, 1989). from pairspawnsboth withandwithout streakers(n = 16
The possibility that sperm depletion in males during dates). Usingthisdatabase, both the presence ofa streaker
the daily spawning period causes lower fertilization rates joiningthe pairspawn and theday a spawn wascollected
wasexamined bytestingforanegativecorrelation between
fertilization rate and the orderofcollection ofspawns for
aspecific maleduringa spawningperiod. Onlysequences 100- o
with at least seven samples were used for this purpose. w 8 o o
o
For statistical analysis, fertilization rate data were an- 5 80 e o o
gularlytransformed beforeapplyingparametric statistical
techniques. In cases where the sea-condition ranking or Co
the spawning-order ranking was used, non-parametric ' 60-
statistics were employed. N
Results oS 40-
12345
A total of 180 collections ofeggs were taken over 19
days. Of the 180 collections, 102 (57%) had the mini- 20
mum
of 20 eggs and were used in the analysis of fertil-
ization rate. Themean fertilization rateofthesecollections
was88.1%(median = 87.3%)(Fig. 1).Thenumberofeggs Sea conditions (rank)
collected from a spawn varied from to 536 (median
= Figure 2. The relationshipbetween seaconditions(rank score)and
= 2656..55);. fTohrectohlrleeectcioonnstrowlitchollaetctlieoansst h2a0d e0g-g4s,egmgse,diiamn- m>e2a0negfgesr.tilization rate on a day for all spawns without streakers with
FERTILIZATION RATE IN H.4LIC1IOERES 235
had a significant effect on fertilization rate (Table I). TableI
tSiplaizwantisonwirtahtestorneaakedrsayha(draanngeav=er-ag4etoof61.86%%) hciogmhperarfeerd- aTnwdo-cwolalyecmtiioxnedda-ymoodnelfeArtNiOlilza'tAionofrtalheeeffect ofpresenceofstreakers
with pair spawns without streakers.
The identity ofthe male did not appear to affect fer- Dependent variable Fertilization rate(angulartransformation)
tilization rate. On two days collections were made si-
Independent
multaneously from three males on the same back reef
area within 30 meters ofone another. There was a sig-
nificant effect of date (Fug = 6.41, P = 0.02), but no
significant effect ofmale identity (F2.it = 0.56, P = 0.58).
There was also no evidence that fertilization rate for
spawns by a male changed in any systematic way during
the spawning period. Over the course ofthe study, there
werethreesequencesof7-10spawnssuccessfullycollected
from a male. These sequences occurred over a period of
33-48 min. The order in the spawning sequence was not
=
significantly correlated with fertilization rate (rs -0.25
to0.64, Quailed] > 0.1 in all cases). Thus, the fertilization
rate data does not provide evidence for sperm depletion
by TP male H. bivattatus.
Discussion
Fertilization rate has been a previously ignored aspect
of the reproductive biology of tropical reef fishes. The
resultthatlowerfertilization ratesoccurondaysofrougher
waterconditionsand highercurrent velocity isconsistent
with a theoretical treatment oftheeffect ofturbulence on
fertilization rate(Dennyand Shibata. 1989)andwithem-
pirical data for marine invertebrates (Pennington, 1985;
Levitan, 1989) and another marine fish (Petersen el
a/.,1992).
A difference in fertilization rate was also observed be-
tween pair spawns and pair spawns joined by streaking
males. The most likely cause oftheincrease in fertilization
rate in spawns with streakers is a higher total amount of
sperm released with the addition ofanother male. How-
ever, thisresult could alsobecaused by someother factor
associated with streaking. Forexample, aslowerspawning
ascentoralowerspawningrush height mightbeassociated
both with increased fertilization rates and with a higher
probability of additional males joining the spawn. Al-
though these alternatives are possible, 1 will assume that
the most straightforward interpretation is correct, that
streakersdirectlyincreasefertilization ratein Halichoeres
bivattatus. Ifso, the result that streaking increases fertil-
ization rate indicates that some ofthe assumptions that
have been made about reproduction in reef fishes have
beenviolatedandthatfertilizationratemustbeconsidered
asapotentialselective forceinshapingspawningbehavior.
If females attempt to maximize fertilization rates in
their spawns, irrespective ofthe identity ofthe male that
fertilizes the eggs, then these data provide evidence for a
conflict ofinterests between the TP territorial male and
236 C. W. PETERSEN
havior. Hesitations by females before beginning the types of males, it is clear that our assumptions of how
spawning rush in tropical reef fishes could be a female mating success is translated into reproductive success in
mating tactic to increase streaking rate, and not a female these fishes needs to be reappraised.
response to imperfect positioning by the TP male prior These results suggest two avenues for future research
to spawning or a tactic to reduce streaking rate. This hy- in the reproductive biology oftropical reef fishes. First,
pothesis needs to be tested with comparative data, in- variation in female mate choice, spawning site selection,
cludingspecieswithand without streakingmales. Females and temporal patternsofspawningmaybeat least in part
in H. hivattatus often hesitate several times before ulti- a response to variation in fertilization rate. Second, the
mately spawning with a TP male (pers. obs.). spawning behavior of both males and females needs to
Despitethepossibleadvantagesoffemalespawningbe- bereconsideredasapotential tactictoincreaseindividual
havior that increases fertilization rate, these modifications fertilization rates.
may be highly constrained by predation pressure during
spawning. Behaviors that may increase the chance ofa
Acknowledgments
nearby maleseeingandjoiningthespawn, such asslowing
down the spawning rush, may also increase the vulnera- This study was supported by NOAA grant NURC
bility ofthe female to predators. In two instances during 90-1 and aSmithsonian Tropical Research Institute post-
this study, a lizardfish (Synodus sp.) struck at the fish in doctoral fellowship to the author. I would like to thank
a pair spawn at the apex ofthe spawning rush. Although Sarah Cohen, Helen Hess, and Amy Sewell for help with
neither attempt was successful, predators may limit the data collection and Helen Hess and two anonymous re-
advantages offemale behaviors that could increase their viewersformakinghelpful commentsonthe manuscript.
visibility to streakers but at the same time make them
more vulnerable to predators. Thresher (1984) has noted
Literature Cited
that many of the observations of reef-fish being preyed
upon have occurred during spawning. Baird,T.A.1988. Abdominalwindowsinstraight-tailedrazorfish,.\Y-
In another Caribbean wrasse, Thalassoma bifasciatum, richlhvs maninicensis: an unusual female sex character in a poly-
fertilization rates do not differ between pair spawns and gynousfish. Copeia 1988: 496-499.
groupspawns(Petersen ct a!., 1992), despitean estimated vancdheonicBeeragned,tEh.ePm.a,tFi.ngWecronsetsruosf,paenridphRe.raRl.mWaalrense.rA.ni19m8.9.BehaFve.ma3l8e:
80-foldincrease in sperm released ingroupspawnsrelative 875-884.
to pair spawns. Group spawns typically involve at least Denny, M. \\., and M. F. Shibata. 1989. Consequences ofsurf-zone
five males, and may produce higher levels ofturbulence turbulence for settlement and external fertilization. Am. Nat. 134:
and water mixing, reducingthe concentration ofgametes 859-889.
Gross, M. R.,and E. L.Charnov. 1980. Alternativemalelifehistories
faster than in pair spawns. Group spawns may also lack in bluegill sunfishes. Proc. Nat. Acad. Set. U.S.A. 76:6937-6940.
theclosejuxtaposition ofamaleand femaleduringgamete Hoffman,S.G. 1985. Effectsofsizeandsexonthesocialorganization
release that exists between the pair-spawning male and of reef-associated hogfishes, Bodianus sp. Environ. Bio/ Fish. 14:
female with or without streaking (Petersen et ai, 1992). 185-197.
Thus, fertilization rate comparisons between different Kuwamura,T. 1984. SocialstructureoftheprotoynousfishLabroides
dimtdialus Publ. SewMar. Biol Lab 29: 117-177.
typesofspawns maydepend both on the numberofmales Lejeune, P. 1987. Theeffect oflocal stockdensity on social behavior
in the spawn and the type ofspawn. This difference for and sex change in the Mediterranean labrid Comjulis. Environ
spawns with streaking and group spawns compared with Bid Fish. 18: 135-141.
pairspawnsshouldbetested forotherspeciestodetermine Levitan,D.R. 1989. Lifehistoryandpopulationconsequencesofbody
sizeregulation inthesea urchinDiademaantillarnm Phillipi. Ph.D.
its generality.
Dissertation UniversityofDelaware, Newark.
Intransforming male matingsuccessinto reproductive Moyer,J.T., and V. Yogo. 1982. The lek-like matingsystem ofHal-
success, researchers have had to rely on approximations ichivrex melanochir (Pisces: Labridae) at Miyake-jima, Japan. Z.
ofhow fertilizations are divided among males in cases of Ticrpsychol. 60: 209-226.
sperm competition. The approach has varied from as- Nemtzov, S.C. 1985. Socialcontrolofsexchange in the Red Sea ra-
signing all ofthe reproductive success to the male using zorfishXyrichtyspentadactylus(Teleostei, Labridae). Environ. Biol.
f-'uh. 14: 199-211.
the alternative reproductive pathway due to his higher Pennington, J. T. 1985. The ecology offertilization ofechinoid eggs:
sperm production and proximity tothe female (Grossand theconsequencesofspermdilution,adultaggregation,andsynchro-
Charnov, 1980)todividingthefertilizationevenlyamong nousspawning. Biol. Bull. 169:417-430.
all males in the spawn (Warner and Hoffman 1980a, b; Petersen,C.W. 1987. Reproductivebehaviourandgenderallocation
Petersen 1987, 1990). In all ofthese studies, fertilization in Serranusfasciatus, a hermaphroditic reeffish.Anim. Behav. 35:
1601-1614.
rate was assumed constant. This study shows that addi- Petersen, C. \V. 1990. The relationships among population density.
tional malesalter the fertilization rate, and, although it is individual size, mating tactics, and reproductive success in a her-
not clear how fertilizations are divided among the two maphroditic fish, Serranusfasciatus. Behaviour 113: 57-80.
FERTILIZATION RATE IN HALICHOERES 237
Petersen,C.W. 1991. Sexallocationinhermaphroditicseahasses.Am. wrasses (Labridae: Halichoeres spp.) offthe coast ofFlorida. Mar.
Nat. (in press). Biol. 53: 161-172.
Petersen,C. W., R. R.Warner,S.Cohen,H.C.Hess,andA.T.Sewell. Thresher,R.E.1984. ReproductioninReefFishes.T.F.H.Publications.
1992. Variation in pelagic fertilization success: implications for NeptuneCity. NJ.
productionestimates,matechoice,andthespatialandtemporaldis- Tribble, G. W. 1982. Social organization, patterns ofsexuality, and
tribution ofmating. Ecology(in press). behaviorofthewrasse Corisdorsomacu/alaat Miyake-jima, Japan.
Robertson, D. R. 1972. Socialcontrolofsexreversalincoral-reeffish. Environ. Biol. Fish. 1: 29-38.
Science111: 1007-1009. Victor,B.C. 1987. ThematingsystemoftheCaribbeanrosyrazorfish,
Robertson, D. R. 1981. The social and mating systems oftwo lahrid Xyrichthysmartinicensis. Bull. Mar. Sci. 40: 152-160.
fishes. Halichoeres maculipinna and H. garnoti. offthe Caribbean Warner. R.R. 1984a. Matingbehaviorand hermaphroditism incoral
coastofPanama. Mar. Biol. 64: 327-340. reeffishes.Am. Sci. 72: 128-136.
Robertson,D.R.,andJ.H.Choat.1974. Protogynoushermaphroditism Warner, R. R. 1984b. Deferred reproduction asa response to sexual
andsocialsystemsinlabridfishes. ProceedingsSecondInternational selectioninacoralreeffish:atestofthelifehistoricalconsequences.
Symposium CoralReefs1: 217-225. Evolution38: 148-162.
Robertson, D. R.,andS.G. Hoffman. 1977. Therolesoffemalemate Warner,R.R. 1987. Femalechoiceofsitesversusmatesinacoralreef
choiceandpredation inthematingsystemsofsometropicallabroid fish, Thalassoma bifasciatiim. Anim. Behav. 35: 1470-1478.
fishes. Z Tierpsychol. 45: 298-320. Warner,R.R. 1988. Traditionallyofmating-sitepreferencesinacoral-
Robertson,D.R.,andR.R.Warner. 1978. Sexualpatternsinthelabroid reeffish, \atiire335: 719-721.
fishesofthewesternCaribbean. II:Theparrotfishes(Scaridae).Smith. Warner, R. R., and S.G. Hoffman. 1980a. Local populationsizeasa
Contrih. Zoo/. 255: 1-26. determinantofmatingsystemandsexualcompositionintwotropical
Ross, R. M. 1986. Social organization and matingsystem ofthe Ha- reeffishes(Thalassomaspp.)Evolution34: 508-518.
waiian reef fish Thalassoma dupcrrey (Labridae). Pp. 794-802 in Warner,R. R.,andS.G. Hoffman. 1980b. Populationdensityandthe
Indo-Padfic Fish Biology: Proceedings oftheSecondInternational economicsofterritorialdefenseinacoralreeffish.Ecology61: 772-
ConferenceonIndo-PacificFishes. T. Uyeno, R. Arai. T. Taniuchi. 780.
and K. Matsuura, eds. Ichthyological SocietyofJapan. Tokyo. Warner,R.R.,andD.R.Robertson.1978. Sexualpatternsinthelabroid
Shapiro,D.V. 1989. Sexchangeasanalternativelife-historystyle. Pp. fishes ofthe western Caribbean, I: The wrasses (Labridae). Smith.
177-195mAlternativeLife-historyStylesinAnimals, M.N. Burton, Contrih. Zool. 254: 1-27.
ed. Kluwer Publishers, Dordrecht,The Netherlands. Warner,R.R.,D.R.Robertson,andE.G.Leigh,Jr.1975. Sexchange
Thresher, R. E. 1979. Social behavior and ecology oftwo sympatric andsexual selection. Science 190: 633-638.